For engineering teams
Development teams adopting Claude Code, Copilot, Cursor or agents, who want the speed without shipping code nobody read.
Your engineers have agentic coding tools and are using them. Pull requests are bigger and arrive faster. Reviewers are drowning, tests are thin, and nobody can say whether the code base is getting better or just getting larger. You want to keep the speed and get the evidence.
Typical climb
Rung 2 to rung 4
Where most people in this group start, and where the work gets them. Your own starting rung might differ, which is the first thing we find out.
What is not working
- Write-only mode: code goes in, nobody reads it, and the review queue is the new bottleneck.
- No evals, so a prompt tweak or a model bump can quietly break behaviour.
- Agent set-up is per developer and inconsistent across repos.
- Guardrails live inside prompts, where they are unreliable.
What we do together
- Set up the agent per repo properly: instructions files, skills, MCP servers, and the conventions that make output consistent.
- Run a habits diagnostic on your repositories before and after, so you can see whether reading and testing are keeping pace with generation.
- Put evaluations in CI: golden examples, a scorecard, and a build that goes red on regression.
- Move guardrails out of prompts and into deterministic checks after the model.
- Apply the promotion ladder to agents: which tasks are shadow, advisory, assisted or autonomous, and what evidence moves each one up.
What you leave with
- Consistent agent set-up across your repositories.
- An eval suite that gates changes, seeded from your own failures.
- A measured before-and-after on reading, testing and review habits.
- A written ladder for your agents, with promotion criteria.
Format
- A two-day workshop on your code base, remote or onsite in the UK.
- Optional fractional support afterwards, by the day, to embed it.
Rates
What it costs, and the pilot offer
The first three organisations get 40 per cent off any team, engineering or leadership engagement. In return I ask for a written testimonial and a short case study I can publish. The individual rates are already the pilot rates.
| Engagement | Remote | Onsite (UK) |
|---|---|---|
| Two-day adoption workshopOn your code base. Agent set-up, habits diagnostic, evals in CI, the agent ladder. | £5,800 | £6,800 |
| Fractional supportEmbedding it afterwards, by the day. | £750 per day | £850 per day |
Tell me which rung you are on.
A short message is enough. Say who you are, what you are trying to do, and whether you want remote or onsite. I reply within two working days.